Abstract
Background: "Leukemia stressors” refer to extrinsic factors that may facilitate clonal progression. We previously reported that smoking initiates unique leukemogenic signatures in acute myelogenous leukemia (AML) characterized by a higher proportion of C>A and G>A single base substitutions (SBS) [Hoang. ASH. 2021]. These G>A SBS are known to be particularly prevalent in CpG islands due to cytosine methylation. However, toxic smoking metabolites could potentially induce specific genomic signatures in patients with AML that resemble COSMIC SBS4 in lung cancer [i.e. enrichment in C>A transversion], and spontaneous C>T SBS1 signatures within specific genomic clusters or nodes. In this study, we investigate how these SBS1/SBS4 signatures distribute within hot-spot regions and evaluate whether G>A mutations (mut), commonly observed in methylated CpG islands, are differentially expressed in AML patients who are older and/or are smokers. In our study, these hot-spot regions or nodes were referenced as a genomic pool 1 (GP1).
Methods: After IRB approval, the incidence of specific SBS densities were investigated in vulnerable hot-spot regions affected by exogenous [SBS4] and endogenous [SBS1] mutagenesis. Next generation sequencing (NGS) was used in patients to obtain data regarding specific mutation types and locations. In known gene exons, we examined genomic coordinates corresponding to individual mut. COSMIC browser allowed genomic coordinate examination to "aggregate” mut within putative "vulnerable exon neighborhoods". "Hot-spot neighborhoods” were described as genomic pool 1 (GP1). Descriptive and inferential statistics were used to analyze data, which was obtained from a single institution.
Results: 80 patients (pt) diagnosed with hematopoietic malignancies with available next generation sequencing (NGS) were selected for analysis. 73/80 (91%) were AML pt. The median age was 63 years (y) (range, 22-88). 42% and 64% of pt were smokers and older than 63 y, respectively. With respect to specific hot spots or nodes, AML pt showed increased G>A mutagenicity in GP1 cluster 1 (Fig 1). Specifically, for G>A, smoking increased the rate from 18.4% to 37.5% (p=0.01) (Fig 1 A and B). Age > 63 y led to an increased mutagenicity rate of 17.1% to 30.5% (p=0.03) (Fig 1 C and D). A synergistic effect in mutagenicity was observed when combining age > 63 y plus smoking (10.3% to 41.9%, p=0.01) (Fig 1. E and F). In regards to hotspots in specific genes, smoking has the ability to increase NRAS (G>A), P53 (G>A) and SRSF2 (C>A) frequencies within GP1. Interestingly, smoking increased C>A mutagenicity in the SRSF2 gene in GP1 from 50% vs 82%.
Conclusions: Our data suggests that aging and smoking globally and locally erode the AML genome by potentially accelerating cytosine methylation and thus G>A mutagenicity, most likely in CpG islands within genes like NRAS and P53. These genes likely contain hot spots aggregated in GP1 that, when mutated, can de-stabilize the genome and increase the risk of AML. Importantly, we observed similar COSMIC SBS4 substitutions seen in lung cancer associated with C>A transversions in older AML pt who had a history of smoking, which is specifically found in the SRSF2 gene. Overall, our data advances our understanding on how risk factors for AML affect the human AML genome. These have strong implications for AML prevention, pathogenesis and therapy. Future studies should be aimed at identifying other intrinsic and extrinsic risk factors for AML seeking to explore their specific mutational patterns and locations within the human genome.
Disclosures
No relevant conflicts of interest to declare.
